Exploring The Terrorist Acitivities And Anti-terrorism Strategies Based On Open-Source Intelligence

Terrorism is an outstanding publichazard in the world of today, and its threat to world peace and stability is severe and imminent. For China, The most severe and realistic terrorist threat is from the “East Turkistan” separatist forces. It is theoretically and physically vital to study the “East Turkistan” terrorist activities from the perspective of system science.This paper aims to solve practical problems in anti-terrorism and it is based on internet information. By using knowledge in fields of social network, complex network, textual analysis and multi-agent simulation, the paper deeply and systematically carries out research on personnel and organizational network in “East Turkistan” terrorism and separatist activities, temporal and spatial distribution characteristics of the activities, terrorist information communication and coping strategies, modeling and simulation of the riots and so on. The main research work and innovatory results of the paper are as follow:1. It proposed a new analysis method for the research on terrorist activities, namely, “Six Element” analysis method. A variety of sub-networks are constructed according to the correlation among the six elements—people, organization, time, location, manner and event. These sub-networks are assessed through centrality analysis, spectrum analysis, cohesive subgroup analysis, spatial correlation analysis and invulnerability analysis; and therefore, the characteristics and laws of terrorist activities are revealed from different levels and perspectives.2. The “East Turkistan” terrorist personnel and organization research based on man-made network. The paper studied “East Turkistan”activities in the past 60 years, applied “Six Element” analysis method, and constructed “terrorist-terrorist” one-mode network and “terrorist-organization” bipartite network. In the one-mode network, centrality analysis and spectrum analysis methods are used to describe the importance of each vertex. The weighted extreme optimization(WEO) algorithm is also applied to detect communities in the one-mode network. In the terrorist-terrorist organization bipartite network, the author adopt centrality analysis and clustering analysis based on the original bipartite network to calculate the importance of each vertex, and finally apply the edge clustering coefficient algorithm to detect the communities. Finally, the control and combate strategies of the “East Turkistan” terrorist activity network are studied based on bootstrap percolation.3. Spatial-temporal characteristics and manners analysis of terrorist activities based on man-made network. This paper applied terrorist activities “Six Element” analysis method to study spatial-temporal characteristics in terrorist activities. Three basic networks, namely, “event-time”, “event-location” and “time-location” are constructed. By analyzing these basic and derivative networks and adopting GIS techniques, the spatial-temporal characteristics of these activities are revealed in different viewpoints. Furthermore, the main manners of these activities are studied.4. “East Turkistan” ethnic separatist activities study on auto-made network. Firstly, we collect the data of the East Turkistan separatist activities from World Wide Web by utilizing web crawlers, adopt text analysis techniques to extract the fundamental elements of “East Turkistan” separatist activities including person, organization, time and place, and then construct multi-mode networks according to the relations between concepts. Later, we apply the approach of dimensional decomposition to separate the multi-mode networks into single vertex sub-networks and bipartite sub-networks, distinguish the importance of different types of nodes by measuring the centrality of each sub-network, and integrate the centrality of each sub-network to form the importance composite index(ICI). To further identify the core nodes in the network, we then adopt the k-shell decomposition method to directly decompose multi-mode networks. In the end, we find that these results produced by approaches above are well agreed with actual results.5. Spreading and optimal intervention policies based on Multi-Agents. This paper builds a heterogenous multi-agent model based on the Net Logo software by using a BA scale-free network model to capture the social structure. Adopting both random strategy and preferencestrategy, wesimulate the spreading of terrorism in the network and intervention policy of government to study the influencing factors for the spreading and the optimal intervention strategy. The results indicate that the effects of spreading of terrorism news depend on both network structure and individuals’ conformity, but have little to do with the spreading strategy of terrorism organizations. The optimal intervention policy mainly depends on individuals’ conformity and intervention of the government, namely, a stronger conformity increases the efficient time of the intervention policy by the government. Furthermore, a larger intervention scope promotes a better result both for random intervention strategy and preference intervention strategy.6. Modeling and simulation of “the 5th, July riot” based on multi-agents. We adopt the multi-agents method to model on “the 5th, July riot”, including four types of agents such as participating terrorists, potential terrorists, civilians and police. We present a safety factor to describe the extent of the environment where the agent is located, and the parameter affects the state of agent so as to crate a nonlinear feedback among the behavior, result and environment of the agent, completing a integral complex adaptive system. We analyze the sensitivity of initial agent to reflect individual deceiving phenomenon and the broken window effect in reality. At last, we propose the coping measures to the riots based on our experimental results.